When computers were first developed, they mainly stored conventional numeric or character data. But today, computers are increasingly being used to store, access, and manipulate not only numeric and character data but also video images, still images, audio data, or a combination of types of data, which can require a large amount of storage. Because of their size, these types of data are often stored in databases in a special way in untyped objects called blobs (binary large objects). The blobs are untyped, meaning that the database system does not know the format of the data. Thus, blobs are typically stored in a database with only two attributes: a data value and a data length, which is the length of the data value.
The lack of formatting information causes a problem when attempting to convert blob data between languages when the blob data type is used to store character data. Modem computers are capable of supporting multiple languages with different character sets and converting their data between them. For example, the English language uses different characters than does the Russian, Chinese, Japanese, or Korean languages, and users would like to be able to convert their data between the different languages using a conversion program. Unfortunately, since blob data is untyped, the conversion program is unable to convert the blob data. In order to take advantage of blob data and to provide support for multiple languages, what is needed is a technique for converting blob data objects between languages.